Sunday, 1 October 2017

The Earth sciences no longer need the publishers for publishing



Manuscript servers are buzzing around our ears, as the Dutch say.

In physics it is common to put manuscripts on the ArXiv server (pronounced: Archive server). A large part of these manuscripts are later send to a scientific journal for peer review following the traditional scientific quality control system and assessment of the importance of studies.

This speeds up the dissemination of scientific studies and can promote informal peer review before the formal peer review. Manuscripts do not have copyrights yet, so this also makes the research available to all without pay-walls. Expecting the manuscripts to be published on paper in a journal later, ArXiv is called a pre-print server. In these modern times I prefer manuscript server.

The manuscript gets a time stamp, a pre-print server can thus be used to claim precedence. Although the date of publication is traditionally used for this and there are no rules which date is most important. Pre-print servers can also give the manuscript a Digital Object Identifier (DOI) that can be used to cite it. A problem could be that some journals see a pre-print as prior publication, but I am not aware of any such journals in the atmospheric sciences, if you do please leave a comment below.

ArXiv has a section for atmospheric physics, where I also uploaded some manuscripts as a young clouds researcher. However because most meteorologists did not participate it could not perform the same function as it does in physics; I never got any feedback based on these manuscripts. When ArXiv made uploading manuscripts harder to get rid of submissions by retire engineers, I stopped and just put the manuscripts on my homepage.

Three manuscript archives

Maybe the culture will now change and more scientists participate with three new initiatives for manuscript servers for the Earth sciences. All three follow a different concept.

This August a digital archive started for Paleontology (paleorXiv, twitter). If I see it correctly they already have 33 manuscripts. (Only a part of them are climate related.) This archive builds on the open source preprint server of the Open Science Framework (OSF) of the non-profit Center for Open Science. The OSF is a platform for the entire scientific workflow from idea, to coding and collaboration to publishing. Also other groups are welcome to make a pre-print archive using their servers and software.

Two initiatives have just started for all of the Earth sciences. One grassroots initiative (EarthArXiv) and one by AGU/Wiley (ESSOAr).

EarthArXiv will also be based on the open source solution of the Open Science Framework. It is not up yet, but I presume it will look a lot like paleorXiv. It seems to catch on with about 600 twitter listeners and about 100 volunteers in just a few days. They are working on a logo (requirements, competition). Most logos show the globe; I would include the study of other planets in the Earth sciences.

The American Geophysical Union (AGU) has announced plans for an Earth and Space Science Open Archive (ESSOAr), which should be up and running early next year. They plan to be able to show a demo at the AGU's fall meeting in December.

The topic would thus be somewhat different due to the inclusion of space science and they will also permanently archive posters presented at conferences. That sounds really useful; now every conference designs their own solution and the posters and presentations are often lost after some time when the homepage goes down. EarthArXiv unfortunately seems to be against hosting posters. ESSOAr would also make it easy to transfer the manuscripts to (AGU?) journals.

A range of other academic societies are on the "advisory board" of ESSOAr, including EGU. ESSOAr will be based on proprietary software of the scientific publisher Wiley. Proprietary software is a problem for something that should function for as close to an eternity as possible. Not only Wiley, but also the AUG itself are major scientific publishers. They are not Elsevier, but this quickly leads to conflicts of interest. It would be better to have an independent initiative.

There need not be any conflict between the two "duelling" (according to Nature) servers. The manuscripts are open access and I presume they will have an API that makes it possible to mirror manuscripts of one server on the other. The editors could then remove the ones they do not see as fitting to their standards (or not waste their time). Beyond esoteric (WUWT & Co.) nonsense, I would prefer not to have much standards, that is the idea of a manuscript server.



Paul Voosen of Nature magazine wonders whether: "researchers working in more sensitive areas of the geosciences, such as climate science, will embrace posting their work prior to peer review." I see no problem there. There is nothing climate scientists can do to pacify the American culture war, we should thus do our job as well as possible and my impression is that climatology is easily in the better half of the Open Science movement.

I love to complain about it, but my impression is that sharing data is more common in the atmospheric sciences than average. This could well be because it is more important because data is needed from all over the world. The World meteorological Organization was one of the first global organizations set up to coordinate this. The European Geophysical Union (EGU) has open review journals for more than 15 years. The initial publication in a "discussion" journal is similar to putting your manuscript on a pre-print server. Many of the contributions to the upcoming FORCE2017 conference on Research Communication and e-Scholarship that mention a topic are about climate science.

The road to Open Access

A manuscript server is one step on the way to an Open Access publishing future. This would make articles better accessible to researchers and the public who paid for it.

Open Access would break the monopoly given to scientific publishers by copyright laws. An author looking for a journal to publish his work can compare price and service. But a reader typically needs to read one specific article and then has to deal with a publishers with monopoly power. This has led to monopolistic profits and commercial publishers that have lost touch with their customers, the scientific community. That Elsevier has a profit margin of "only" 36 percent thus seems to be mismanagement, it should be close to a 100 percent.



ArXiv shows that publishing a manuscripts costs less than a dollar per article. Software to support the peer review can be rented for 10 dollar per article (see also: Episciences.org and Open Journal Systems). Writing the article and reviewing it is done for free by the scientific community. Most editors are also scientists working for free, sometimes the editor in chief gets some secretarial support, some money for a student help. Typesetting by journals is highly annoying as they often add errors doing so. Typesetting is easily done by a scientist, especially using Latex, but also with a Word template. That scientists pay thousands of dollars per article is not related to the incurred costs, but due to monopoly brand power.

Publishers that serve the community, articles that everyone can read and less funding wasted on publishing is a desirable goal, but it is hard to get there because the barriers to entry are large. Scientists want to publish in journals with a good reputation and if the journals are not Open Access with a broad circulation. This makes starting a new journal hard, even if a new journal does a much better job at a much lower price, it will start with no reputation and without a reputation it will not get manuscripts to prove its worth.

To make it easier to get from the current situation to an Open Access future, I propose the concept of Grassroot Scientific Publishing. Starting a new journal should be as easy as starting a blog: Make an account, give the journal name and select a lay-out. Finished, start reviewing.

To overcome the problem that initially no one will submit manuscripts a grassroots journal can start with reviewing already published articles. This is not wasted time because we can do a much better job communicating the strength and weakness as well as the importance of an article than we do now, where the only information we have on the importance is the journal in which it is published. We can categorise and rank them. We can have all articles of one field in the same journal, no longer scattered around in many different journals.

Even without replacing traditional journals, such a grassroots journal would provide a valuable service to its scientific community.

To explain the idea and get feedback on how to make it better I have started a new grassroots publishing blog:
Once this kind of journals is established and has shown it provides superior quality assurance and information, there is no longer any need for pay-wall journals and we can just review the articles on manuscript servers.

Related reading

Paul Voosen in Nature: Dueling preprint servers coming for the geosciences

AGU: ESSOAr Frequently Asked Questions

The Guardian, long read: Is the staggeringly profitable business of scientific publishing bad for science?

If you are on twitter, do show support and join EarthArXiv

Three cheers for gatekeeping

Peer review helps fringe ideas gain credibility

Grassroots scientific publishing


* Photo Clare Night 2 by Paolo Antonio Gonella is used under a Creative Commons Attribution 2.0 Generic (CC BY 2.0) license.

13 comments:

Victor Venema said...

The discussions increasingly move from the blog comments to social media. (Is there a way to automatically make tweets into comments for blogger, like for Wordpress?)

On twitter a professorial bunny worries more than me about bad manuscripts: "How do you keep the nutters away?". (Note, a US chemistry bunny, knowledgeable about climate, but not a climate scientist.)

A reply I did not think of above, but is also important is that: A few nutters may be useful to prevent press from treating a manuscript server as if it were the scientific literature. You might need a few percent nutters to do this effectively.

It should be clear that manuscripts on a server are not much different from someone putting a manuscript on their own homepage, it has no a-priory credibility. If we start policing which manuscripts are allowed, that could give the wrong impression that there is some credibility to the manuscripts. This would be bad, that is what a real peer review by domain experts is for, a cursory look by the manage of the server is not peer review.

On ArXiv someone has to "endorse" the manuscripts, but this is only to make sure that they are on topic and a real manuscript. This does not check the quality of the manuscript.

Jérémy Anquetin said...

Interesting post, and thanks for the shout-out for PaleorXiv (I am part of the steering committee).

I think what you are proposing is very close to the Peer Community In project (https://peercommunityin.org). They started it for Evolutionary Biology (http://evolbiol.peercommunityin.org/), but now other communities are organizing themselves. I am leading the initiative for the future Peer Community in Paleontology (PCI Paleo).

I would be happy to provide more information if you are interested.

Best,
Jérémy.

Victor Venema said...

Hi Jérémy, welcome and thank you. Great to learn about Peer Community In. I was thinking much in the same direction. I see they just started. Looks like this is the time.

Was thinking of keeping the peer review more traditional (that is, closed at least if the authors prefer). I also expect that many reviewers would prefer to be anonymous. Writing a named review is more work.

I was hoping to add more value than normal reviews by publishing the reviews (but also anonymous reviews allowed) like you do, but also by giving more (quantitative) feedback on the importance of the articles.

My ideal grassroots journal would also keep the reviews up to date (post-publication review). By categorising the articles it would help outsiders to find their way into the literature, as well as by listing the most important papers per category (like normally your supervisor would).

Jérémy Anquetin said...

With PCI the reviews are always public (more transparency in the review process), but the reviewers may choose to remain anonymous if they prefer.

And there is also added value in the review process because the recommender (the member of the community who managed the review process) is asked to write a recommendation detailing why the paper is important. The recommendation can be written by the recommender alone, or jointly by the recommender and the reviewers. Such recommendations have DOIs and can be cited, which can be useful to all (readers, authors, and recommender/reviewers).

The PCI system is based on a environment where the preprint is the unity. Post-publication reviews can always be posted on the preprint server and linked "physically" with the preprint.

Finally, categorising, if needed, can be done by overlay journals or online collections (such as Science Open collections).

But, I am not sure the Open Access publishing future, as you call it, really needs journals at all.

Victor Venema said...

No idea whether you need to call it a journal, but also the open access world would need structure to make sure that all relevant articles get a review and to help scientists find interesting articles.

Just a big database with a bit of up and down voting will not work. Science is a human enterprise and the expertise of the people making the assessments matters.

One article could be listed in multiple grassroots journals. (Is it then still a journal?)

I could also imagine partial endorsements/reviews. I work on homogenisation of climate data, our grassroots homogenisation journal could assess the quality of the homogenisation for an article that also studies climatic changes based on that homogenised data. (Is it then still a journal?) Our assessment on homogenisation could help another journal that is interested in the climate changes found.

Also thought of the terms: a collection, a curation (is a foreigner allowed to redefine words?) and a community journal.

A Science Open collection comes close to what I had seen as the first task, at least if they do more than "just" select and write a short summary. Once it is established that the community quality assessment works, I would hope that people will become willing to also submit manuscripts.

Looks like PCI feels that the peer review of manuscripts is the first step (which can then still be submitted to a "real" journal for the official peer review as long as the community peer review is not yet accepted). I had not thought of doing peer review without "publishing". That is a nice idea to immediately start with interactive peer review and not just writing an assessment of final manuscript/paper.

Stefan Schmeja said...

Just to clarify: Physicists usually don't put manuscripts on arXiv and send them later to journals, but rather the other way round. Most papers are posted on arXiv after they have been submitted to a journal or even after they have been accepted.

Blair Trewin said...

The Journal of Southern Hemisphere Earth Systems Science (formerly the Australian Meteorological and Oceanographic Journal) moved a couple of years ago to a model where authors do their own typesetting. It seems to work OK. (JSHESS is now online-only and open-access, and doesn't have any publication charges, since its costs are fairly minimal apart from the voluntary time of the editors/reviewers).

(Disclosure: I was editor of AMOJ/JSHESS at the time this change was made, although I'm not any more).

Victor Venema said...

Stefan, thanks for the info. That was also what I used to do. I thought the idea was to do it the other way around and get some feedback, maybe even while still working on more details, further tests.

Blair, that confirms my impression that running a journal does not have to cost much. Most of the work is already done for free by us. The representative of the publisher tried to convince me that there is a lot of additional work and investment, but was not able to tell in detail what it was, beyond needing a lot of time to get rid of spam contributions. I would expect a community journal to get no spam contributions because they know the community and a spam contribution has no chance whatsoever.

EliRabett said...


So Eli asks why are there so many nutty papers these days. There is a simple answer, Naomi Oreskes. Before Oreskes' paper there were essentially none in the climate area. Oreskes was practically a dare to the nuts. Cook et al was a double dare, and with the availability of predatory journals and on line publishing they have exploded, to the point that only when one sneaks into a "real" one does any bunny take the time to write a comment: (This paper sucks).

There is tons of crazy out there. Remember that Willard Tony owns the most important climate blog.

Jérémy Anquetin said...

Of course, I am not defending a system without any structure. In my view, the future system should be based on preprints and peer-review (managed by scientists) of these preprints. The journals, or whatever we call them, will come only afterwards and sort out papers into thematics or collections, which will be extremely useful, but they will no longer play an essential role in selecting and peer-reviewing science.

I also agree with you regarding the fact that one paper can be endorsed by several grassroots journals. That's actually possible with the Peer Community in system.

Finally, the objective of PCI is also a system where there is no formal "publishing" (i.e., peer-review of preprints, without subsequent publication). However, the PCI founders also know that any totally disruptive initiative would have difficulties to take roots. One of the advantages of PCI is precisely that you can start using this system and prove that it works, while at the beginning continuing to submit your work to traditional journals. After a time, all actors will realize that these journals do not bring any real added value and that a recommendation by a PCI (or any other peer-reviewing/recommending platform) is enough and is worth as much as any validation by a traditional journal.

Elio Campitelli said...

Yes! I really hope these arXives take off and becomes the standard way of publishing research.
As a young researcher starting my career I'm trying my best to learn the good practices of open science and make them an integral part of my workflow as early as possible.

Nikola Vitas said...

There are two interesting stories from the solar physics. First, in 2010 a new journal was founded by the Max Planck Institute for Solar System Research in Goettingen - Living Reviews in Solar Physics. The journal is fully electronic with possibility of updates by authors and comments by the audience. A truly wonderful concept. A scientific journal made for scientists, entirely by scientists - no wonder it gained IF close to 18 in only 4 years. However, in 2015 it was sold to Springer!!!

And something else, a bit different: some months ago a series of on-line solar physics seminars has been started - the idea is that 10+ European institutes organize common seminar talk once every two weeks. The talk is given at one of the institutes and broad-casted to others through the Zoom platform. It works amazingly well. Actually, it's so good that one has to wonder about the future of conferences (or, at least, how many conference do we actually need).

Victor Venema said...

That sounds like a wonderful journal. I started thinking about this kind of publishing when I noticed some weaknesses in a much-cited paper of mine. I now wrote a blog post about these weakness and told many people at conferences, but it would have been much nicer if this would have been stored together with the article. This problems are not that big that if warrants the huge overhead of an official comment, but I think the readers would be happy to know.

A pity it was sold to Springer. Another reason to make it a Creative Commons community work so that others can continue, which decreases the price Springer would be willing to pay.

I do not think conferences will go away fully, but once you know people you can do a lot via video. I have always wondered why it took so long in times of skype. For smaller groups it is already quite common to have one kick-off meeting in person and do the rest via the internet.

In case of big meetings like AGU or EGU you could have satellite meetings on other continents, greatly reducing intercontinental flights.